AIO Loaders have been removed as of v12. If you use AIO models, you can easily add a checkpoint loader to replace the existing loaders and connect them accordingly.
The default Load Diffusion Model loader looks for your Krea 2 Turbo model in the models/diffusion_models folder.
This workflow has the following functions/features:
Text to Image
Multiple Lora Loading
Wildcards
Upscale
Detailers (removed on v12)
Save Image with full resource metadata in CivitAI readable format
These will work with Anima Turbo models, but you'll have to figure out the settings.
Support Policy
These workflows are provided as-is and without guaranteed support.
I do not provide support for third-party applications, modified versions, redistributed copies, reposts, bundles, or installations obtained from unofficial sources.
Inclusion of these workflows in another project does not imply my endorsement, involvement, or responsibility for that project.
Disclaimer
Responsibility for generated content falls on the user of these workflows. I do not control how people prompt or what they use.
Downloading/installation of ComfyUI and/or any related custom nodes for these workflows are at the installer's own risk.
Description
Krea 2 Turbo v13 changes since v12b
Fixed pixelation when bypassing the second latent pass. The latent upscale and red sampler now share the 2nd Pass / Upscale group. Bypassing it sends the blue sampler’s latent directly to VAE Decode at its original size.
Added optional Pixel Upscale after VAE Decode. Use Lanczos on its own or enable Upscale w/ Model first. The included 4× UltraSharpV2 Lite model produces an intermediate image, and Upscale (Initial Image Size) by sets the final dimensions relative to the decoded image. Pixel Upscale is bypassed by default.
Updated SeedVR2 and output routing. SeedVR2 Final Cleanup is enabled by default and runs after the optional pixel upscale. Image Comparer A shows the final output, while B and Image Result show the Krea VAE decode before pixel upscale and SeedVR2. Image Saver records the dimensions of the final image.
Added aspect-ratio and megapixel controls. Resolution Selector replaces the SDXL preset latent node and defaults to 2:3 portrait at 1 MP.
Added a shared Negative Prompt. Both Krea samplers now use CFG Guider at
1.0, with the negative prompt wired to both samplers and Image Saver. The field starts blank.Updated the prompt reformatter and text encoder. Replaced TextGenerate with TextGenerateV2, revised the system prompt, raised the token limit to
4096, and connected the seed. Prompt Reformat is bypassed by default. The saved text encoder changed from Qwen3-VL 4B FP8 scaled to Qwen3-VL 4B BF16.Updated metadata and workflow notes. Image Saver now reads CFG from the blue guider. The notes explain the red refinement pass, pixel upscale sizing, and the manual step-count setting. Set Steps to
0when bypassing the red pass and back to4for the default refinement.Updated saved defaults. Comfy Kitchen Attention starts enabled. Both LoRA loaders have preset entries saved with their toggles off.